1. Field of the Invention
The present invention is directed to methods and compositions of modified variants of diphtheria toxin (hereinafter “DT”) that reduce binding to vascular endothelium or vascular endothelial cells, therefore, reduce the incidence of Vascular Leak Syndrome (hereinafter “VLS”).
2. Background Information
Vascular Leak Syndrome is primarily observed in patients receiving protein fusion toxin or recombinant cytokine therapy. VLS can manifest as hypoalbuminemia, weight gain, pulmonary edema and hypotension. In some patients receiving immunotoxins and fusion toxins, myalgia and rhabdomyolysis result from VLS as a function of fluid accumulation in the muscle tissue or the cerebral microvasculature [Smallshaw et al., Nat Biotechnol. 21(4):387-91 (2003)]. VLS has occurred in patients treated with immunotoxins containing ricin A chain, saporin, pseudomonas exotoxin A and DT. All of the clinical testing on the utility of targeted toxins, immunotoxins and recombinant cytokines reported that VLS and VLS-like effects were observed in the treatment population. VLS occurred in approximately 30% of patients treated with DAB389IL-2 [(Foss et al., Clin Lymphoma 1(4):298-302 (2001), Figgitt et al., Am J Clin Dermatol, 1(1):67-72 (2000)]. DAB389IL-2, is interchangeable referred to in this application as DT387-IL2, is a protein fusion toxin comprised of the catalytic (C) and transmembrane (T) domains of DT (the DT toxophore), genetically fused to interleukin 2 (IL-2) as a targeting ligand. [Williams et al., Protein Eng., 1:493-498 (1987); Williams et al., J. Biol. Chem., 265:11885-11889 (1990); Williams et al., J. Biol. Chem., 265 (33):20673-20677, Waters et al., Ann. New York Acad. Sci., 30(636):403-405, (1991); Kiyokawa et al., Protein Engineering, 4(4):463-468 (1991); Murphy et al., In Handbook of Experimental Pharmacology, 145:91-104 (2000)]. VLS has also been observed following the administration of IL-2, growth factors, monoclonal antibodies and traditional chemotherapy. Severe VLS can cause fluid and protein extravasation, edema, decreased tissue perfusion, cessation of therapy and organ failure. [Vitetta et al., Immunology Today, 14:252-259 (1993); Siegall et al., Proc. Natl Acad. Sci., 91(20):9514-9518 (1994); Baluna et al., Int. J. Immunopharmacology, 18(6-7):355-361 (1996); Baluna et al., Immunopharmacology, 37(2-3): 117-132 (1997); Bascon, Immunopharmacology, 39(3):255 (1998)].
Reduction or elimination of VLS as a side effect would represent a significant advancement as it would improve the “risk benefit ratio” of protein therapeutics, and in particular, the immunotoxin and fusion toxin subclasses of protein therapeutics. (Baluna et al., Int. J. Immunopharmacology, 18(6-7):355-361 (1996); Baluna et al., Immunopharmacology, 37(No. 2-3):117-132 (1997); Bascon, Immunopharmacology, 39(3): 255 (1998). The ability to develop fusion proteins, single chain molecules comprised of a cytotoxin and unique targeting domain (scfv antibodies in the case of immunotoxins) could facilitate the development of the therapeutic agents for autoimmune diseases, such as rheumatoid arthritis and psoriasis transplant rejection and other non-malignant medical indications. (Chaudhary et al., Proc. Natl. Acad. Sci. USA, 87(23):9491-9494 (1990); Frankel et al., In Clinical Applications of Immunotoxins Scientific Publishing Services, Charleston S.C., (1997), Knechtle et al., Transplantation, 15(63):1-6 (1997); Knechtle et al., Surgery, 124(2): 438-446 (1998); LeMaistre, Clin. Lymphoma, 1:S37-40 (2000); Martin et al., J. Am. Acad. Dermatol, 45(6):871-881, 2001)). DAB389IL-2 (ONTAK) is currently the only FDA approved protein fusion toxin and employs a DT toxophore and the cytokine interleukin 2 (IL-2) to target IL-2 receptor bearing cells and is approved for the treatment of cutaneous T-cell lymphoma (CTCL). (Figgitt et al., Am. J. Clin. Dermatol, 1(1):67-72 (2000); Foss, Clin. Lymphoma, 1(4):298-302 (2001); Murphy et al., In Bacterial Toxins: Methods and Protocols, Hoist O, ed, Humana Press, Totowa, N.J., pp. 89-100 (2000)). A number of other toxophores, most notably ricin toxin and pseudomonas exotoxin A, have been employed in developing both immuntoxins and fusion toxins; however, these molecules have not successfully completed clinical trials and all exhibit VLS as a pronounced side effect (Kreitman, Adv. Pharmacol, 28:193-219 (1994); Puri et al., Cancer Research, 61:5660-5662 (1996); Pastan, Biochim Biophys Acta., 24:1333(2):C1-6 (1997); Frankel et al., Supra (1997); Kreitman et al., Current Opin. Inves. Drugs, 2(9):1282-1293 (2001)).
VLS arises from protein-mediated damage to the vascular endothelium. In the case of recombinant proteins, immunotoxins and fusion toxins, the damage is initiated by the interaction between therapeutic proteins and vascular endothelial cells. Lindstrom et al. provided evidence that ricin toxin A had direct cytotoxic effects on human umbilical vein vascular epithelial cells but that these effects were not mediated by fibronectin (Lindstrom et al., Blood, 90(6):2323-34 (1997); Lindstrom et al., Methods Mol. Biol, 166:125-35 (2001)). Baluna et al. postulated that the interaction disrupts fibronectin mediated cell-to-cell interactions resulting in the breakdown of vascular integrity, and Baluna further suggested that in the toxin ricin, the interaction is mediated by a conserved three amino acid motif, (x)D(y), where x is L, I, G or V and y is V, L or S (Baluna et al., Int. J. Immunopharmacology, 18(6-7):355-361 (1996); Baluna et al., Proc. Natl. Acad. Sci. USA, 30:96(7):3957-3962, (1999); Baluna et al., Exp Cell Res., 58(2):417-24 (2000)). It was reported that one of the VLS motifs found in ricin toxin, the ‘LDV’ motif, essentially mimics the activity of a subdomain of fibronectin which is required for binding to the integrin receptor. Integrins mediate cell-to-cell and cell-to-extracellular matrix interactions (ECM). Integrins function as receptors for a variety of cell surface and extracellular matrix proteins including fibronectin, laminin, vitronectin, collagen, osteospondin, thrombospondin and von Willebrand factor. Integrins play a significant role in the development and maintenance of vasculature and influence endothelial cell adhesiveness during angiogenesis. Further, it is reported that the ricin ‘LDV’ motif can be found in a rotavirus coat protein, and this motif is important for cell binding and entry by the virus. (Coulson, et al., Proc. Natl. Acad. Sci. USA, 94(10):5389-5494 (1997)). Thus, it appears to be a direct link between endothelial cell adhesion, vascular stability and the VLS motifs which mediate ricin binding to human vascular endothelial cells (HUVECs) and vascular leak.
Mutant dgRTAs were constructed in which this motif was removed by conservative amino acid substitution, and these mutants illustrated fewer VLS effects in a mouse model (Smallshaw et al. Nat Biotechnol., 21(4):387-91 (2003)). However, the majority of these constructs yielded dgRTA mutants that were not as cytotoxic as wild type ricin toxin, suggesting that significant and functionally critical structural changes in the ricin toxophore resulted from the mutations. It should also be noted that no evidence was provided to suggest that the motifs in dgRTA mediated HUVEC interactions and VLS in any other protein. Studies revealed that the majority of the mutant dgRTAs were much less effective toxophores and no evidence was provided to suggest that fusion toxins could be assembled using these variant toxophores.
DT is composed of three domains: the catalytic domain; transmembrane domain; and the receptor binding domain (Choe et al. Nature, 357:216-222 (1992)). Native DT is targeted to cells that express heparin binding epidermal growth factor-like receptors (Naglish et al., Cell, 69:1051-1061 (1992)). The first generation targeted toxins were initially developed by chemically cross-linking novel targeting ligands to toxins such as DT or mutants of DT deficient in cell binding (e.g. CRM45). (Cawley, Cell 22:563-570 (1980); Bacha et al., Proc. Soc. Exp. Biol. Med., 181(1):131-138 (1986); Bacha et al., Endocrinology, 113(3):1072-1076 (1983); Bacha et al., J. Biol. Chem., 258(3):1565-1570 (1983)). The native cell binding domain or a cross-linked ligand that directs the DT toxophore to receptors on a specific class of receptor-bearing cells must possess intact catalytic and translocation domains. (Cawley et al., Cell, 22:563-570 (1980); vanderSpek et al., J. Biol. Chem., 5:268(16):12077-12082 (1993); vanderSpek et al., J. Biol Chem., 7(8):985-989 (1994); vanderSpek et al., J. Biol Chem., 7(8)985-989 (1994); Rosconi, J. Biol Chem., 10;277(19):16517-161278 (2002)). These domains are critical for delivery and intoxification of the targeted cell following receptor internalization (Greenfield et al, Science, 238(4826)536-539 (1987)). Once the toxin, toxin conjugate or fusion toxin has bound to the cell surface receptor the cell internalizes the toxin bound receptor via endocytic vesicles. As the vesicles are processed they become acidified and the translocation domain of the DT toxophore undergoes a structural reorganization which inserts the 9 transmembrane segments of the toxin into the membrane of the endocytic vesicle. This event triggers the formation of a productive pore through which the catalytic domain of the toxin is threaded. Once translocated the catalytic domain which possess the ADP-ribosyltransferase activity is released into the cytosol of the targeted cell where it is free to poison translation thus effecting the death of the cell (reviewed in vanderSpek et al., Methods in Molecular Biology, Bacterial Toxins: methods and Protocols, 145:89-99, Humana press, Totowa, N.J., (2000)).
Chemical cross-linking or conjugation results in a variety of molecular species representing the reaction products, and typically only a small fraction of these products are catalytically and biologically active. In order to be biologically active, the reaction products must be conjugated in manner that does not interfere with the innate structure and activity of the catalytic and translocation domains in the toxophore. Resolution of the active or highly active species from the inactive species is not always feasible as the reaction products often possess similar biophysical characteristics, including for example size, charge density and relative hydrophobicity. It is noteworthy that isolation of large amounts of pure clinical grade active product from chemically crosslinked toxins is not typically economically feasible for the production of pharmaceutical grade product for clinical trials and subsequent introduction to clinical marketplace. To circumvent this issue, a genetic DT-based protein fusion toxin in which the native DT receptor-binding domain was genetically replaced with melanocyte-stimulating hormone as a surrogate receptor-targeting domain was created (Murphy et al, PNAS, 83:8258-8262 (1986)). This approach was used with human IL-2 as a surrogate targeting ligand to create DAB486IL-2 that was specifically cytotoxic only to those cells that expressed the high-affinity form of the IL-2 receptor (Williams et al., Protein Eng., 1:493-498 (1987)). Subsequent studies of DAB486IL-2 indicated that truncation of 97 amino acids from the DT portion of the molecule resulted in a more stable, more cytotoxic version of the DL-2 receptor targeted toxin, DAB389IL-2 (Williams et al., J. Biol Chem., 265:11885-889 (1990)). The original constructs (the 486 forms) still possessed a portion of the native DT cell binding domain. The DAB389 amino acid residue version contains the C and T domains of DT with the DT portion of the fusion protein ending in a random coil between the T domain and the relative receptor binding domain. A number of other targeting ligands have since been genetically fused to this DT toxophore, DAB389. (vanderSpek et al., Methods in Molecular Biology, Bacterial Toxins:Methods and Protocols., 145:89-99, Humana Press, Totowa, N.J. (2000)). Similar approaches have now been employed with other bacterial proteins and genetic fusion toxins are often easier to produce and purify.